Which metrics reveal hallucination in a support bot?
No single metric is enough. Track grounded-answer rate, unsupported-claim rate, contradiction rate against approved policy, and escalation quality. Together they show whether the bot answers from evidence, invents details, conflicts with the source of truth, or hands uncertain cases to a person.
Measure these by intent and policy area rather than only across the whole bot. A strong average can hide a dangerous failure in refunds, billing, eligibility, or account security.
Complaint volume is useful but late. The best operational signal is a reviewed sample of real conversations linked back to the exact retrieved evidence and the answer the model produced.
- Category
- Product
- Platform
- Web